Klaus Ma

icon indicating a website link http://klaus1982.cn icon indicating an email [email protected]

icon company Nvidia icon location Beijing China icon location Founder of Volcano & Flame. Tech Lead of CNCF TAG-Runtime and CNCF Batch WG.

Results 89 comments of Klaus Ma

Request to add more metrics for spark job running on k8s scheduled by volcano

+1, that's great!!!

Suggest to add a field to queue.spec indicating queue open or closed

Can we delete all jobs fristly, and then delete the queue; we can handle those actions in `QueueController`, similar to namespace.

`elect`/`reserve`/`reservation` system's locked nodes are not actually locked, and other pods schedule on them

> Thanks for your feedback. Let me take a check. Before this feature was released, it has been tested locally for several different scenarios. /assign AFAIK, we're going to replace...

Add GPU Numbers Predicates

Let's create a PR for design doc :)

RFVE: Support GPU topology

/cc @carmark @Jeffwan

RFVE: Support GPU topology

@carmark , thanks very much for the info :) I'd like to build something based on that example :)

RFVE: Support GPU topology

/kind rfve

Cromwell workflows not working

/cc @Thor-wl

support maping queue to node group

and it's better to have a CRD named `ResourceGroup`, so we know how many nodes will be used in a specific queue :)

support maping queue to node group

> > and it's better to have a CRD named `ResourceGroup`, so we know how many nodes will be used in a specific queue :) > > Does it aims...